ADAPTATION OF A PARALLEL ALGORITHM FOR THE ACCELERATED CALCULATION OF THE DISCRETE FOURIER TRANSFORM FOR REALIZATION ON VECTOR COMPUTERS.

The minimization of conversion time is one of the fundamental requirements when solving many problems in digital signal processing in discrete orthogonal bases. In a previous paper, the author describes a new parallel discrete Fourier transformation algorithm which separates with a parametrically extended vector of the simultaneously processed data. As a result of this, a considerable reduction in the number of iterations of the computing process was obtained compared with 'classical' fast Fourier transform algorithms, and consequently, the computer time required was used. However, the main drawback of this algorithm is the need for multipass summation of simultaneously applied data. An improved approach is considered in this paper. It is pointed out that the approach used to construct the proposed version of the parallel algorithm can be extended to the Vinograd algorithm, polynomial transformations, and other algorithms for fast discrete transformations in different orthogonal bases.